TMEM52B Suppression Promotes Cancer Cell Survival and Invasion

Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 https://doi.org/10.1186/s13046-021-01828-7

RESEARCH Open Access TMEM52B suppression promotes cancer cell survival and invasion through modulating E-cadherin stability and EGFR activity Yunhee Lee1,2†, Dongjoon Ko1,3†, Junghwa Yoon1, Younghoon Lee2 and Semi Kim1,3*

Abstract Background: TMEM52B is a novel gene broadly expressed in a variety of normal human tissues. However, the biological function of TMEM52B expression in cancer is largely unknown. Methods: The effects of TMEM52B on tumor growth and metastasis were investigated in vitro and in vivo, and the underlying biological and molecular mechanisms involved in this process were evaluated. Clinical datasets from KmPlotter and The Cancer Genome Atlas (TCGA) were analyzed in relation to TMEM52B expression and function. Results: Suppression of TMEM52B in colon cancer cells promoted cancer cell epithelial-mesenchymal transition (EMT), invasion, and survival in vitro. Similarly, in vivo studies showed increased tumor growth and circulating tumor cell survival (early metastasis). ERK1/2, JNK, and AKT signaling pathways were involved in TMEM52B suppression- induced invasiveness and cell survival. TMEM52B suppression promoted activation and internalization of epidermal growth factor receptor (EGFR) with enhanced downstream signaling activity, leading to enhanced cell survival and invasion. In addition, TMEM52B suppression reduced E-cadherin stability, likely due to a reduced association between it and E-cadherin, which led to enhanced β-catenin transcriptional activity. Concomitantly, TMEM52B suppression promoted generation of soluble E-cadherin fragments, contributing to the activation of EGFR. Clinical data showed that high TMEM52B expression correlated with increased patient survival in multiple types of cancer, including breast, lung, kidney, and rectal cancers, and suggested a correlation between TMEM52B and E-cadherin. Conclusions: These findings suggest that TMEM52B is a novel modulator of the interplay between E-cadherin and EGFR. It is possible that TMEM52B functions as a tumor-suppressor that could potentially be used as a novel prognostic marker for cancer. Keywords: TMEM52B, Cell survival, Invasion, EGFR, E-cadherin, Prognostic marker

* Correspondence: [email protected] †Yunhee Lee and Dongjoon Ko contributed equally to this work. 1Immunotherapy Research Center, Korea Research Institute of Bioscience and Biotechnology, Daejon, South Korea 3Department of Functional Genomics, Korea University of Science and Technology, Daejon, South Korea Full list of author information is available at the end of the article

© The Author(s). 2021 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data. Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 2 of 20

Background C12orf59 or FLJ31166) was first cloned in 2002 [9] and Metastasis is the leading cause of cancer-related deaths has been recently shown to be expressed at the mRNA in most cancer types. A metastatic cascade has several level in kidney (https://www.genecards.org/). TMEM52B elements, including local invasion, intravasation, cancer mRNA isoform 1 (NM_153022; 2789 bp) encodes a 163 cell survival in the circulation, extravasation, microme- amino acid protein, and isoform 2 (NM_001079815; tastasis, and metastatic colonization [1]. Early in cancer 2563 bp) differs from it in the 5′ UTR, containing an al- metastasis, epithelial tumor cells are activated to invade ternate exon in the 5′ coding region. Isoform 2 also ini- surrounding stromal tissue through the epithelial- tiates translation from an alternate upstream start mesenchymal transition (EMT) process. EMT also plays codon, and encodes a distinct N-terminal, 183 amino important roles in next steps such as intravasation, can- acid protein. The TMEM52B gene shares homology with cer cell survival in the circulation, and micrometastasis. other species, including mouse, rat, frog, and chicken. During EMT, cells undergo morphological and molecu- Recently, broad TMEM52B expression was reported in lar changes to acquire a more mesenchymal phenotype normal human tissues, with high expression in kidney; [2–5]. A hallmark of EMT is the functional loss of E- however, its expression in a panel of cancer cell lines cadherin, a well-known metastasis suppressor [5]. E- was not detectable [10]. In that study, decreased cadherin mediates cell-cell adhesion through Ca2+- TMEM52B expression was correlated with poor survival dependent homophilic interactions, thereby suppressing and von Hippel-Lindau mutations in renal cell carcin- cell migration and dissemination. In addition, E- oma patients [10], suggesting that decreased TMEM52B cadherin anchors β-catenin to the cell membrane as part expression may promote renal carcinogenesis. of a complex to prevent β-catenin from entering the nu- However, general biological or cancer-biology func- cleus and inducing the expression of EMT-related genes, tions for TMEM52B expression have yet to be deter- thereby inhibiting EMT and suppressing cancer metasta- mined. Here we report that suppression of TMEM52B sis [5]. Downregulation of E-cadherin is usually medi- in cancer cells increases cell survival and invasion ated by E-cadherin transcriptional repressors/EMT- in vitro, and enhances tumor growth and early metasta- inducing transcription factors and is also mediated at sis in vivo. TMEM52B suppression enhanced EGFR acti- the post-translational level [5, 6]. Furthermore, soluble vation and the downstream MAPK and AKT signaling E-cadherin fragment, formed by cleavage of E-cadherin, pathways, leading to cancer cell invasion and survival. has been reported to contribute to tumor progression by TMEM52B suppression mediated the shedding of extra- promoting invasion and metastasis [6]. cellular E-cadherin and contributed to EGFR activation. Epidermal growth factor receptor (EGFR) plays key These results suggest that TMEM52B is a novel modula- roles in essential cellular functions, including prolifera- tor of EGFR and E-cadherin, with tumor/metastasis-sup- tion, migration, and survival. EGFR is also an important pressing activity. Thus, a reduction in TMEM52B driver of tumorigenesis mostly in glioblastoma, as well expression may contribute to tumor progression. as lung, breast, and colon cancers. Inappropriate activation of EGFR through gene amplification, mutation, ab- Materials and methods normal protein overexpression, ligand overproduction, Cell lines and dysregulated intracellular trafficking plays an im- Human embryonic kidney 293E (HEK293E), SW480, portant role in cancers [7]. EGF-to-EGFR binding at the Colo205, HCT-15, HCT-116, HT-29, SW620, Caco-2 cell surface induces receptor dimerization and activation (colon cancer), KATOIII, MKN28, MKN45 (gastric can- of its tyrosine kinase domain and downstream signaling cer), SKOV3, and IGROV1 (ovarian cancer) cell lines events, including both PI3K/AKT and RAS/RAF/MEK were purchased from the American Type Culture Col- pathways, which are critical for controlling proliferation, lection (ATCC; Manassas, VA, USA). SW480sub cell migration, and survival. Concomitantly, the receptor- was a subclone isolated from SW480 cells in our labora- ligand complexes are also internalized into endosomal tory, which displayed elevated cell-matrix adhesion cap- compartments and routed to lysosomes for degradation, acity compared to SW480 cells. SNU-216, SNU-638, thereby terminating receptor activation and downstream SNU-668, and SNU-719 (gastric cancer) cell lines were signaling, or are sorted for recycling. However, internal- purchased from the Korean Cell Line Bank (KCLB; ized EGF-EGFR complexes can continue signaling within Seoul, Korea). HEK293E cells were maintained in early endosomes and multivesicular bodies until their DMEM with 10% fetal bovine serum (FBS) at 37 °C in disassembly and degradation. Therefore, EGFR endo- 5% CO2. SW480, HCT-116, HT-29, Colo205, HCT-15, cytosis can result in either positive or negative effects on SW620, KATOIII, MKN28, MKN45, SKOV3, and signaling and tumorigenesis [7, 8]. IGROV1 cells were maintained in RPMI1640 with 10% Transmembrane protein 52B (TMEM52B; gene ID, FBS at 37 °C in 5% CO2. Caco-2 cells were maintained in 120939; chromosome location, 12p13.2; also termed MEM with 10% FBS at 37 °C in 5% CO2. Cell were Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 3 of 20

checked for mycoplasma and their identities were con- ATCTCTGGTATATAT-3′), TMEM52B isoform 2 (5′- firmed using STR-PCR analysis. GGACGAGAUGUGAGGAAAATT-3′) or EGFR (5′- CGCAAAGUGUGUAACGGAAUATT-3′) using Lipo- Plasmid and cDNA constructs fectamine 2000 for 48 h prior to analysis. cDNAs encoding two isoforms of wild-type full-length TMEM52B-specific siRNA (5′-GGGTACATCTCTGG human TMEM52B were amplified by PCR with primer TATATAT-3′) and scrambled siRNA (5′-ACTATATC set (Forward primers for both vectors are 5′- AACATC TAGTTAGGTGCTG-3′) were subcloned into the TCGAGGCCGCCATGTCGTGGCGGCCTC-3′ for iso- pLKO.1 lentiviral shRNA vector (Addgene, Cambridge, form 1 and 5′- AACATCTCGAGGCCGCCATGGGAG MA, USA) digested with AgeI/EcoRI to generate pLKO.1- TCCGAGTTCAT-3′ for isoform 2. Reverse primers for shTMEM52B and pLKO.1-shControl, respectively, ac- both isoforms are 5′-AATTCAAGCTTGTTCCAA cording to the manufacturer’s instructions. TMEM52B- GAGTCAACTATTCG-3′ for pcDNA3.1-myc and 5′- specific siRNA was also introduced into BglII/HindIII- CTTCGAATTCTGTTCCAAGAGTCAACTATTCG-3′ digested pSUPER vector (OligoEngine, Seattle, WA). The for pEGFP-N1) using cDNA from SW480 cells. The EGFR-expressing vector was provided by Dr. M.-C. Hung PCR product was digested with XhoI/HindIII and XhoI/ [11]. Human E-cadherin-expressing vector was previously EcoRI and then cloned into the expression vectors described [12]. Cells were transfected using Lipofectamine pcDNA3.1-myc and pEGFP-N1, respectively. The 2000 or electroporation (Invitrogen). At 48 h after trans- pcDNA3.1-myc was generated by inserting stop codon fection, cells were lysed for immunoblot analysis or har- after myc tag into pcDNA3.1-myc-his vector. For flow vested for further analysis. A cDNA fragment encoding cytometry analysis, cDNA fragments encoding two iso- the human E-cadherin extracellular domain was sub- forms of TMEM52B (without the putative signal peptide cloned into the pCMV-Fc-myc vector [13]. The resulting sequence (amino acid residues 1–24) in the case of iso- recombinant soluble E-cadherin-Fc (sE-cad-Fc) fusion form 2) with a myc tag at the N-terminus were amplified protein expression plasmid was transfected into HEK293E by PCR and then subcloned into pcDNA3.1 vector/ cells. At 48 h after transfection, the medium was changed XhoI-HindIII. PCR primer set used was as follows; for- to serum-free medium. Conditioned medium was col- ward primers are 5′-CTCGAGGCCGCCATGGAA lected and then subjected to affinity chromatography on a CAAAAACTCATCTCAGAAGAGGATCTGTCGTG Protein A excellose column (Bioprogen, Daejon, Korea) to GCGGCCTCAGCCC-3′ for isoform 1 and 5′-CTCGAG obtain purified sE-cad-Fc fusion protein. GCCGCCATGGAACAAAAACTCATCTCAGAAGA GGATCTGGAGGAAAACTGTGGTAAT-3′ for iso- Generation of stable cell lines form 2 and reverse primer for both isoforms is 5′- To generate lentiviruses, pLKO.1-shTMEM52B or AAGCTTTCAGTTCCAAGAGTCAACTATTCG-3′. pLKO.1-shControl was co-transfected with Lentiviral Packaging Mix (Sigma, St Louis, MO) into Lenti-X-293 Reverse transcription-polymerase chain reaction (PCR) T cells (Clontech, Mountain View, CA) using Lipofecta- Total RNA was isolated using TRIzol (Invitrogen, Carls- mine 2000, and virus-containing supernatants were har- bad, CA), and cDNA was synthesized using reverse tran- vested and concentrated at 48 h post-transfection. HCT- scriptase (Bioneer, Daejon, Korea). Real-time quantitative 15 cells were transduced with the lentiviruses for 9 h in PCR was performed using SYBR Green (PKT, Seoul, the presence of polybrene (5 μg/ml) and were subse- Korea) on a Rotor-Gene 6000 real-time rotary analyzer quently selected with puromycin (15 μg/ml) for a week (Corbett, San Francisco, CA) with TMEM52B (both iso- to establish stable clones. forms)-specific primers (5′- AGTGACCGTCATTGCTTT CG-3′ and 5′-GGCCAAACACCGACTGCAG − 3′), iso- Immunoblot analysis form 1-specific primers (5′-TGGTCAGGCAACAGGA Whole-cell lysates were prepared using RIPA buffer as TGAA-3′ and 5′-ACAGCAGCAGCGGAAGCAC-3′), described previously [14] and analyzed using the follow- isoform 2-specific primers (5′- GTGGTCAGGCAATG ing primary antibodies: anti-Histone H3, anti-β-actin TTCAGG-3′ and 5′-ACAGCAGCAGCGGAAGCAC-3′), and anti-GAPDH (Santa Cruz Biotechnology, Santa and GAPDH-specific primers (5′-CATGACCACAGTCC Cruz, CA); anti-myc (Upstate Biotechnology, Lake Pla- ATGCCAT-3′ and 5′-AAGGCCATGCCAGTGAGC cid, NY); anti-α-catenin, anti-β-catenin, anti-E-cadherin, TTC-3′) with an annealing temperature of 61 °C. and anti-integrin α5 (BD Biosciences, San Jose, CA); anti-α smooth muscle actin (αSMA) and anti-vimentin Transfection of siRNAs, shRNA vectors, and expression (Sigma); anti-phospho-c-Jun N-terminal kinase (JNK) vectors (T183/Y185), anti-phospho-extracellular signal-regulated Cells were transfected with siRNA specific to TMEM52B kinase 1/2 (ERK1/2), anti-ERK1/2, anti-phospho-AKT both isoforms (Dharmacon, Lafayette, CO; 5′-GGGTAC (S473), anti-AKT, anti-cyclin D1, anti-JNK, anti- Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 4 of 20

phospho-c-Jun(S63), anit-phospho-ATF-2(T71), anti-c- calculating stained area relative to the total area using Jun, anti-ATF-2, anti-Slug, anti-bcl-2, anti-survivin, anti- ImageJ software. cleaved caspase-3, anti-PARP, anti-phospho- EGFR(Y1068), anti-EGFR, anti-phospho-β-catenin(S33, Immunofluorescence staining 37/T41), anti-phospho-β-catenin(S675), and anti- Cells were plated on coverslips and were transfected with phospho-β-catenin(S552) (Cell Signaling, Danvers, MA); shRNA for 48 h. Cells were then fixed for 15 min in 10% and anti-E-cadherin (R&D Systems, Minneapolis, MN). formalin and permeabilized for 5 min with 0.3% Triton X- To generate anti-TMEM52B antibodies, rabbits were 100. After blocking in 10% normal goat serum, cells were immunized with a peptide (KQLPPTEKESTRIVDSWN), stained using an anti-E-cadherin antibody (R&D Systems; which was derived from the common region of both 1:50) followed by an Alexa546-conjugated secondary anti- TMEM52B isoforms (amino acid residues from 166 to body. EGF-treated transfected cells were fixed/perme- 183 of isoform 2), conjugated with KLH. After boostings, abilized for 10 min in methanol. After blocking in 1% sera were collected to analyze anti-TMEM52B immu- BSA, the cells were incubated with an anti-EEA1 (Cell Sig- noactivity by ELISA (A-frontier, Seoul, Korea) and im- naling; 1:100) or anti-LAMP2 antibody (Abcam, Cam- munoblotting using cell lysates. The sera were further bridge, UK; 1:100) followed by a FITC-conjugated purified by affinity chromatography using protein A/G- secondary antibody and then with an Alexa555- agarose (Millipore, Billerica, MA). conjugated anti-EGFR antibody (Cell Signaling; 1:100) Cells were counterstained with 4,6-diamidino-2-phenylin- Subcellular fractionation dole (DAPI; Sigma) to visualize nuclei. Mounted samples Subcellular fractions were prepared using the Compart- were visualized with a confocal microscope (LSM 510 mental Protein Extraction Kit (Millipore) according to the META; Carl Zeiss, Jena, Germany). Intensity of E- manufacturer’s instructions. The cytosolic fraction, nu- cadherin staining was quantitated by ImageJ software. clear fraction, plasma membrane-enriched fraction, and Quantitative co-localization analysis of EGFR and EEA1 cytoskeletal fraction were analyzed by immunoblotting. or LAMP2 was performed by calculating Pearson’scorrel- ation coefficient (r) using ImageJ software. Co-immunoprecipitation HEK293E cells were transfected with pcDNA3.1-myc- Promoter reporter assay TMEM52B and a vector expressing E-cadherin using Li- For transfection, cells were seeded into 6-well plates at a pofectamine 2000. Two days after transfection, cells density of 2 × 105 cells/well and incubated for 24 h. Cells were lysed with co-immunoprecipitation buffer (10 mM were then transfected with 2 μg of reporter plasmids Tris pH 7.4, 150 mM NaCl, 0.56 mM EGTA, 1% Triton DNA and 1.8 μg of the TMEM52B expression vector X-100, 0.5% NP-40) supplemented with protease inhibi- using Lipofectamine 2000. At 48 h post-transfection, tor (Complete, Roche). Lysates were centrifuged for 20 firefly luciferase activity was measured using a Dual- min at 10000 g and the resulting supernatant was pre- luciferase reporter assay system (Promega, Southampton, cleared by incubation with protein A/G–agarose for 2 h UK). The transfection efficiency was normalized by at 4 °C. The precleared supernatant was immunoprecipi- measuring Renilla luciferase activity, encoded by the co- tated using anti-myc at 4 °C for 16 h. The protein com- transfected Renilla luciferase vector (pRL-TK). The AP-1 plexes were collected by incubation with protein A/G– cis-element reporter plasmids were purchased from agarose for 2 h at 4 °C and then washed four times with Stratagene (La Jolla, CA). β-Catenin-responsive firefly lu- co-immunoprecipitation buffer. The protein complexes ciferase reporter plasmid TOPFlash and the negative were eluted by boiling in sodium dodecyl sulfate sample control FOPFlash were purchased from Millipore. The buffer and analyzed by immunoblotting with anti-myc E-cadherin promoter (− 308/+ 41) reporter plasmid was and anti-E-cadherin. previously described [15].

Invasion assay Anchorage-independent soft agar assay Invasion assays were performed as described previously Cells were seeded at a density of 1 × 103 cells/well in 6- [15]. Cells were plated in serum-free medium on Trans- well tissue culture plates in 0.4% agar (Sigma) over a well inserts (Corning, NY) coated with 25 μg of Matrigel. 0.6% agar feed layer. Cells were allowed to grow at 37 °C The underside of the insert was pre-coated with 2 μgof in 5% CO2 for 13 days, and the number of resulting col- collagen type I (Sigma). After incubation for 48 h at onies was counted per well. 37 °C in 5% CO2, inserts were fixed with 10% formalin and stained with 2% crystal violet. The number of cells Analysis of cell survival and proliferation that had invaded was counted in five representative (× Cell survival under suspension culture conditions was 100) fields per insert or invasiveness was determined by determined. Briefly, cells were seeded into 96-well plates Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 5 of 20

with an Ultra-Low Attachment Surface (Corning) at a For CTC survival analysis (early metastasis model), density of 1.5–2×104 cells/well and incubated for 3 or HCT-15 stable cells (TMEM52B-suppressed cells and 4 days. Cells were then incubated with WST reagent control cells) were injected via the tail vein into nude (Ez-Cytox; Dogenbio, Seoul, Korea; one-tenth of the mice (5 × 106 cells per mouse; n = 5 per group). Mice medium volume) and the amount of formazan dye were sacrificed 24 h after injection and total DNA was formed was determined by measuring absorbance at 450 isolated from the lungs as previously reported [17, 18]. nm using a spectrophotometric microplate reader (BMG To quantify survival and early seeding (arrest) of CTCs, LABTECH GmbH, Ortenber, Germany). human tumor cell contents present in mouse lungs were Cells were seeded into 96-well plates at a density of determined as previously reported [18, 19]. Briefly, real- 3×103 cells/well and incubated for 48 h. Cell prolifera- time qPCR analysis of human prostaglandin E receptor 2 tion was determined by the colorimetric WST assay as (PTGER2) genomic DNA was performed with a PTGE described above. Proliferation rates were also estimated R2-specific primer pair (5′-TACCTGCAGCTGTACG by the growth rate quotient (GRQ) as calculated by the CCAC-3′ and 5′-GCCAGGAGAATGAGGTGGTC-3′) equation: GRQ = (N –N0)/t ×1/ N0, where N is the final and a human PTGER2-specific probe (FAM 5′-TGCT cell number, N0 is the initial cell number and t is the GCTTCTCATTGTCTCG-3′ TAMRA) using Quanti- time elapsed between the two counting [16]. Tect Probe PCR Master Mix (Qiagen, Hilden, Germany) on a Rotor gene Q instrument (Qiagen) according to the manufacturer’s instructions. PCR was performed in trip- Flow cytometry licate with a final volume of 50 μl per reaction using 1 μg To analyze TMEM52B localization at the cell surface, of total genomic DNA as a template. After denaturation HEK293E cells were transfected with expression vectors for 15 min at 95 °C, the reaction was continued for 60 for TMEM52B with a myc tag at the N-terminus for 48 cycles of 94 °C for 15 s and 60 °C for 60 s. In parallel, a h and then stained with PE-conjugated anti-myc anti- standard curve was generated using genomic DNA ex- body (Cell Signaling). Viable propidium iodide-negative tracted from HCT-15 cells and nude mouse lungs. cells were analyzed by flow cytometry. Standard curve samples included serial dilutions of To analyze internalization of EGFR in response to mouse-only, human-only, or human plus mouse mixed EGF treatment, cells were treated with EGF for 5–30 samples of known DNA concentrations. A standard min at 37 °C to allow internalization or maintained at curve with the equation of the linear trend line was de- 4 °C. Cells were transferred to ice-cold buffer to stop the veloped by plotting the mean Threshold Cycle (Ct) occurrence of internalization and stained with FITC- values on the y-axis versus the log amount of human conjugated anti-EGFR antibody (Santa Cruz Biotechnol- genomic DNA on the x-axis. ogy). Propidium iodide-negative cells were analyzed by flow cytometry to analyze residual EGFR on cell surface. Immunohistochemistry and TUNEL assay Formalin-fixed and paraffin-embedded 6 μm-thick tissue Mouse xenograft model and circulating tumor cell (CTC) sections from xenograft tumors were processed for im- survival analysis munohistochemistry analysis as per the standard proto- All animal procedures were performed in accordance col. Sections were stained with anti-Ki67 antibody (SP6; with the guidelines of the Animal Care Committee at Abcam) using the peroxidase technique. The prolifera- the Korea Research Institute of Bioscience and Bio- tion index (%) was determined by calculating the num- technology (KRIBB) and with approval of the bioeth- ber of Ki67-positive cells relative to the total number of ics committee of the KRIBB (KRIBB-AEC-19098). cells, which consisted of at least 1000 cells per field. Ser- Nude mice (BALB/c-nude, 5-week-old females) were ial sections were stained with anti-E-cadherin (1:100 di- obtained from Nara Biotech (Seoul, Korea). HCT-15 lution; R&D Systems) and anti-vimentin (1:100 dilution; stable cells (TMEM52B-suppressed cells and scram- Sigma) and the immune complexes were detected using bled shRNA-transfected control cells) mixed with the Vectastain Elite ABC-HRP kit (Vector Laboratories, Matrigel (1:1) were injected subcutaneously into the Burlingame, CA). The sections were counterstained with right flank of each mouse (5 × 106 cells per mouse; hematoxylin. n = 4 per group). Body weight and tumor volume TUNEL staining was performed to measure apoptosis were measured twice per week. On day 44, mice were in tumor tissue sections using In Situ Apoptosis Detec- sacrificed, and dissected tumor masses were fixed in tion Kit (Abcam) according to the manufacturer’s in- 10% formalin. The tumor volumes were calculated as structions. The apoptosis index (%) was determined by follows: tumor volume = (a × b2) × 1/2, where a was calculating the number of TUNEL-positive cells relative the width at the widest point of the tumor and b was to the total number of cells, which consisted of at least the maximal width perpendicular to a. 1000 cells per field. Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 6 of 20

Analysis of The Cancer Genome Atlas (TCGA) and membranes (Fig. 1c). Neither isoform was present in the kmPlotter data cell nuclei. In addition, HEK293E cells were transiently cBioPortal (www.cbioportal.org)[20, 21] was used to transfected with expression vectors for TMEM52B with analyze Cancer Cell Line Encyclopedia (CCLE; Broad, a myc tag at the N-terminus. Flow cytometry analysis 2019) [22] and TCGA-generated human kidney renal showed that an anti-myc antibody bound to HEK293E clear cell carcinoma data (TCGA, Firehose Legacy). All cells transfected with isoform 2 and isoform 1 (Supple- patient samples where the mRNA expression profiles mentary Fig. S1A and B), confirming that TMEM52B lo- were available were included in our analysis per cancer calized to the plasma membrane. Notably, this antibody study. Calculation of Pearson’s correlation coefficient (r) bound more to HEK293E cells transfected with isoform and survival curve analysis were performed using the 2 than to HEK293E cells transfected with isoform 1. cBioPortal webpage tools. Survival of lung [23], breast Subcellular fractionation and immunoblot analyses [24], liver [25], ovarian [26], rectal, and thyroid [27] can- showed a strong signal intensity in the plasma cer patients within previously published data sets was membrane-enriched fraction, and a moderate signal in- analyzed using KM-plotter (http://kmplot.com). tensity in the cytoplasm fraction from SW480sub cells (Fig. 1d). These results indicate that TMEM52B expres- Statistical analysis sion was mainly present on the plasma membrane. Statistical analyses were performed using the Student’s t- test, one-way ANOVA (with GraphPad Prism 8 (Graph- Suppression of TMEM52B promoted cancer cell invasion Pad Software, CA)), Logrank test (for survival analysis), and survival through activation of MAPKs and AKT and Pearson’s test (for correlation analysis). P < 0.05 was signaling pathways considered statistically significant. To explore the effect of suppressed TMEM52B expression on cell growth, invasion, and survival, SW480, Results SW480sub, and HCT-15 cells were transiently trans- Expression of TMEM52B in cell lines fected with siRNA or shRNA specific to both isoform 1 To explore the expression of TMEM52B in various can- and isoform 2. TMEM52B suppression significantly in- cer cell lines, we performed RT-qPCR analysis. creased invasion of SW480, SW480sub, and HCT-15 TMEM52B mRNA was highly expressed in SW480sub cells by 71, 43, and 51%, respectively (Fig. 2a). Cell sur- (a subclone of SW480, displaying elevated cell-matrix vival under suspension conditions was also increased by adhesion capacity), HCT-15, and Caco-2 colon cancer 67, 138, and 113%, following TMEM52B suppression in cells and was moderately expressed in MKN28 (gastric SW480, SW480sub, and HCT-15 cells, respectively (Fig. cancer), SW480 (colon cancer) cells. The mRNA was lit- 2b). On the other hand, cell proliferation was not sub- tle to no detected in KATOIII, MKN45, SNU-216, SNU- stantially affected by TMEM52B depletion (Supplemen- 638, SNU-668, SNU-719 (gastric cancer), Colo205, tary Fig. S1C and D), whereas anchorage-independent HCT-116, HT-29, SW620 (colon cancer), SKOV3, growth of cells was significantly enhanced by TMEM52B IGROV1 (ovarian cancer), and HEK293E cells (Fig. 1a, suppression (Supplementary Fig. S2A and B). top panel). Notably, semi-quantitative PCR analyses, To evaluate the effect of TMEM52B suppression in vivo, using primers that discriminate between TMEM52B iso- stable HCT-15 cells with suppressed TMEM52B expres- forms, detected isoform 2 mRNA expression in sion were generated and characterized (Supplementary SW480sub, HCT-15, Caco-2, SW480, and MKN28 cells. Fig. S2A–D); consistent with Fig. 2a and b, TMEM52B Isoform 1 mRNA expression was little or no detected suppression enhanced cell survival and invasion, which (Fig. 1a middle and bottom panels), suggesting that iso- was accompanied by enhanced EGFR phosphorylation form 2 was the major transcript. Immunoblot analysis and reduced E-cadherin (see below). These cells were using an antibody raised against a TMEM52B peptide injected subcutaneously into the flanks of nude mice. (KQLPPTEKESTRIVDSWN) showed that the presump- Tumor growth was significantly increased in mouse xeno- tive isoform 2 was substantially expressed in SW480ub grafts with TMEM52B-suppressed cells compared to and HCT-15 cells, and moderately expressed in SW480 those with scrambled shRNA-transfected control cells cells, but not in KATOIII, MKN45, SNU-668, and SNU- (Fig. 2c). Immunostaining of tumor sections from mice 719 cells (Fig. 1b). injected with TMEM52B-suppressed cells showed higher To explore TMEM52B subcellular localization, levels of proliferative Ki67-positive cells compared with HEK293E cells were transiently transfected with GFP- the control-treated mouse tumors (Fig. 2d). Additionally, fused TMEM52B-expressing vectors. Substantive green TUNEL analysis (Fig. 2e) of these same tumor sections fluorescence was detected at the cell membrane edges displayed significantly lower levels of apoptosis compared after isoform 2 transfections. Isoform 1 transfections re- with that of the tumors from mice injected with control sulted in diffused fluorescence in the cytoplasm and cell cells. E-cadherin was substantially detected and vimentin Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 7 of 20

Fig. 1 TMEM52B expression in cell lines. (A) Top, real-time quantitative PCR analysis of TMEM52B in various human cancer cell lines. β-actin was used as an internal control for input RNA. Middle, semi-quantitative RT-PCR analysis of TMEM52B. Densitometric quantification was performed on gel images. Bottom, gel images of semi-quantitative RT-PCR. All determinations were performed in three independent experiments. Values represent mean ± standard deviation (SD). *P < 0.05. (B) Immunoblot analysis of TMEM52B in human cancer cells. GAPDH was used as an internal control. (C) HEK293E cells were transfected with a GFP-fused TMEM52B-expression vector for 48 h before visualizing GFP. Scale bar, 10 μm. (D) Immunoblot analysis of subcellular localization of TMEM52B in SW480sub cells. WCL, whole-cell lysates; C, cytosolic fraction; N, nuclear fraction;M, plasma membrane-enriched fraction; CS, cytoskeletal fraction. GAPDH, Histone H3, EGFR, and vimentin were used as internal controls for the cytosolic, nuclear, and plasma membrane, and cytoskeletal fractions, respectively Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 8 of 20

Fig. 2 Suppression of TMEM52B promotes cancer cell invasion and survival in vitro, and tumor growth and early metastasis in vivo. (A, B) Cells were transfected with siRNA or shRNA specific to TMEM52B for 48 h. (A) Transfected cells (2 × 104 cells/well) were allowed to invade Matrigel on Transwell inserts for 48 h. The number of cells that had invaded was counted in five representative (100×) fields per Transwell insert. Knockdown of TMEM52B was analyzed by semi-quantitative RT-PCR or by immunoblot. (B) To induce anoikis, transfected cells were seeded into 96-well plates with an Ultra-Low Attachment surface and grown for up to 3 days. Cell viability was determined by the colorimetric WST assay. (C) In vivo tumor growth analysis. Stable TMEM52B-suppressed HCT-15 cells (5 × 106 cells/mouse) were injected subcutaneously into the flanks of nude mice (n =4) as described in Materials and Methods. Tumor volume and body weight were measured for 44 days. Tumor volume was calculated using the formula, length × width2/2. Photos of tumor-bearing mice on day 44. (D, E) Ki67 (D) and TUNEL (E) staining of tumor sections from (C) was performed to measure the level of cell proliferation and apoptosis, respectively. Representative images are shown. The proliferative index (%) or apoptosis index (%) was determined by calculating the number of Ki67 or TUNEL-positive cells relative to the total number of cells (at least 1000 cells per field). Five randomly selected fields from tumor sections per mouse were analyzed. Scale bars, 200 μm. (F) Stable HCT-15 cells (5 × 106 cells/mouse) were intravenously injected into nude mice (n = 5) as described in Materials and Methods. At 24 h after injection, the lungs were removed to extract total DNA. Real-time qPCR analysis was performed for human PTGER2 using total DNA extracted from the lungs. The amount of human genomic DNA initially present in a qPCR reaction tube was estimated (top) from the standard curve generated by qPCR using human total DNA extracted from HCT-15 parental cells, which was spiked into mouse total DNA from lungs of nude mice (bottom). All in vitro determinations were performed in three independent experiments. Values represent mean ± SD. *P < 0.05. siCon, scrambled control siRNA, shCon, scrambled control shRNA; siTM, TMEM52B-specific siRNA; shTM, TMEM52B-specific shRNA Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 9 of 20

was not apparently detected in tumor sections of mice (Fig. 3a). Immunoblot analyses showed that TMEM52B injected with TMEM52B-expressing control cells, while suppression induced phosphorylation of JNK (moder- vimentin but not E-cadherin was moderately detected in ately in HCT-15 cells), ERK1/2, and AKT in SW480, tumor sections of mice injected with TMEM52B- SW480sub, and HCT-15 cells. The phosphorylation of suppressed cells (Supplementary Fig. S2E). We also exam- c-Jun and ATF-2, and the expression of cyclin D1, were ined the effect of TMEM52B suppression on the survival also enhanced by TMEM52B knockdown. In addition, of, and subsequent seeding/arrest of, circulating tumor TMEM52B suppression upregulated the mesenchymal cells (CTCs) in vivo. As an early metastasis model, markers, including vimentin (in SW480 and HCT-15 TMEM52B-suppressed stable HCT-15 cells were intra- cells), α smooth muscle actin, integrin α5, and Slug venously injected into nude mice. At 24 h after injection, (moderately). This suppression also downregulated epi- mice were sacrificed, and the lungs immediately removed thelial markers, including E-cadherin (in SW480 and for total DNA extraction to determine tumor cell content. HCT-15 cells; E-cadherin was not apparently detected in Real-time qPCR analysis revealed that the amount of hu- SW480sub cells) and α-catenin (Fig. 3b). TMEM52B man genomic DNA was significantly higher in the lungs knockdown enhanced expression of the anti-apoptotic of mice injected with TMEM52B-suppressed cells than in factors bcl-2 and survivin (in HCT-15 cells). TMEM52B the control-treated mouse lungs (Fig. 2e). This indicates suppression reduced the levels of cleaved caspase-3 and that TMEM52B suppression promotes CTC survival to PARP (in SW480 and HCT-15 cells) under suspension support early metastasis. culture conditions (Fig. 3b). These results indicate that Comparison of the cell morphology revealed enhanced TMEM52B suppression promotes EMT events, leading spreading and formation of membrane ruffle and protru- to cancer cell invasion and survival. Using a reporter sive spots in TMEM52B-suppressed SW480sub cells assay, TMEM52B suppression significantly activated an

Fig. 3 MAPK and AKT signaling pathways are required for cancer cell invasion and survival mediated by TMEM52B suppression. (A) Phase-contrast images of SW480sub control cells and TMEM52B-knockdown cells. Arrowheads indicate representative membrane ruffling or protrusive spot. Number of cells displaying lamellipodia or protrusive spots (more than 3) were counted. Cell spreading area per cell (at least 200 cells) was calculated using ImageJ software. Values represent mean ± SD. *P < 0.05. Scale bar, 25 μm. (B) Cells were transfected with siRNA or shRNA specific to TMEM52B for 48 h prior to lysis for immunoblot analysis or RT-PCR analysis. Transfected cells were incubated for 48 h under suspension culture conditions prior to lysis for immunoblot analysis of cleaved caspase-3 and PARP. (C) Cells were transfected with shRNA specific to TMEM52B for 24 h and then treated with pharmacological inhibitors: PD098059 (5 μM SW480sub and 20 μM HCT-15); SP600125, 15 μM; wortmannin, 10 μMor 25 μM, or 0.1% DMSO as a vehicle control for 24 h before whole-cell lysates were prepared for immunoblotting. (D, E) Cells were transfected with shRNA specific to TMEM52B for 48 h. Invasion (D) and cell survival (E) assays were performed as described in Fig. 2A and B in the presence of pharmacological inhibitors: PD098059 (5 μM SW480sub and 20 μM HCT-15); SP600125, 15 μM; wortmannin, 10 μM, or DMSO. Cell invasion was determined by calculating the cell-stained area relative to the total area using ImageJ software. All determinations were performed in three independent experiments. Values represent mean ± SD. *P < 0.05 compared with shControl + DMSO; §P < 0.05 compared with shTMEM52B + DMSO Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 10 of 20

AP-1 cis-element reporter plasmid (AP-1 reporter), in TMEM52B-specific shRNA vector. Immunoblot analysis SW480 and HCT-15 cells (Supplementary Fig. S3A), in- demonstrated that the TMEM52B suppression-mediated dicating that TMEM52B suppression increases AP-1 phosphorylation of JNK, ERK1/2, c-Jun, and AKT was transcriptional activity. reversed following the suppression of EGFR expression. To determine the role of JNK, ERK1/2, and AKT signal- EGFR suppression also reduced TMEM52B suppression- ing in TMEM52B suppression-mediated cancer cell sur- mediated expression of both integrin α5 and cyclin D1 vival and invasion, SW480sub and HCT-15 cells were (Fig. 4c). TMEM52B suppression-induced invasion (Fig. transiently transfected with TMEM52B-specific shRNA 4d) and cell survival (Fig. 4e) were both significantly re- for 24 h and then treated with the following pharmaco- duced by EGFR knockdown. These results suggest that logical inhibitors for 24 h before analysis: dimethyl sulfox- the suppression of TMEM52B mediated activation of ide (vehicle), PD098059 (a specific MEK/ERK inhibitor), intracellular signaling and promoted invasion and cell SP600125 (a specific JNK inhibitor), or wortmannin (a spe- survival, in an EGFR-dependent manner. cific PI3K/AKT inhibitor). Immunoblot analysis showed signaling suppression by these inhibitors. Inhibition of TMEM52B suppression promoted phosphorylation and MEK/ERK also moderately reduced JNK phosphorylation internalization of EGFR and enhanced downstream in both cells, implying that ERK may be involved in JNK signaling activation. Inhibition of PI3K/AKT also reduced JNK We further explored the effect of TMEM52B suppres- phosphorylation in SW480sub cells, implying that PI3K/ sion on EGFR in response to EGF. TMEM52B- AKT may be involved in JNK activation (Fig. 3c). Inhib- suppressed SW480sub and HCT-15 cells responded to ition of MEK/ERK, JNK, and PI3K/AKT significantly sup- EGF treatment with enhanced and sustained activation pressed both basal and TMEM52B suppression-induced of EGFR. These cells also exhibited activation of down- cell invasion (Fig. 3d) and survival (Fig. 3e). stream signaling mediators ERK1/2, JNK, and AKT In addition, siRNA suppression of only isoform 2 (the (Fig. 5a). Upon ligand binding, EGFR is involved in a major isoform) promoted SW480sub and HCT-15 cell series of trafficking events, which ultimately regulate its invasion and survival to an extent similar to that seen signal amplification, cell survival, and invasiveness. We after suppression of both isoforms (Supplementary Fig. therefore investigated the intracellular distribution of S3B), confirming that isoform 2 is the major isoform EGFR using immunofluorescence staining. In unstimu- and that cellular functions mediated by TMEM52B (sup- lated SW480sub cells, EGFR was predominantly local- pression) are largely attributed to isoform 2. ized to cell surface membranes. Following EGF Together, these results suggest that TMEM52B sup- stimulation, internalized EGFR co-localized with EEA1- pression promotes cancer cell invasion and survival positive early endosomes in both TMEM52B-suppressed through activation of ERK1/2, JNK, and AKT signaling and control cells. TMEM52B-suppressed cells displayed pathways. sustained EGFR localization to early endosomes compared to control cells for up to 30 min (Fig. 5b). In TMEM52B suppression promoted invasion and cell addition, internalized EGFR that co-localized with survival in an EGFR-dependent manner LAMP2-positive lysosomes, was observed less in Next, we explored the upstream signaling responsible TMEM52B-suppressed cells than in control cells at 60 for TMEM52B suppression-mediated signaling pathway min (Supplementary Fig. S4A). Consistently, flow cytom- activation and invasion. We hypothesized that etry analysis showed that residual EGFR levels at the cell TMEM52B may activate several signaling pathways by surface were lower in TMEM52B-suppressed SW480sub modulating specific receptor tyrosine kinase activity due cells, following EGF stimulation, compared to EGFR to TMEM52B being localized to the cell surface and not levels in control cells at both 15 and 30 min (Supple- having any known enzymatic domain. Among the mentary Fig. S4B). screened molecules, EGFR phosphorylation at Tyr1068 In order to examine the functional significance of in- increased in TMEM52B-suppressed HCT-15 and ternalized EGFR signaling (localized to early endosome), SW480sub cells (Fig. 4a). As described below, exogenous dynasore, a cell-permeable dynamin inhibitor, was used EGF-induced phosphorylation of EGFR was also en- to block EGFR endocytosis. As expected, EGFR endo- hanced by TMEM52B suppression. Both basal cell inva- cytosis was inhibited by dynasore in TMEM52B- sion, and that induced by exogenous EGF, were suppressed cells (Supplementary Fig. S4C). Immunoblot enhanced by TMEM52B suppression (Fig. 4b). analysis showed that EGF-induced activations of EGFR, To determine whether EGFR was required for AKT, and ERK1/2 were substantially reduced in TMEM52B suppression-mediated signaling events and TMEM52B-suppressed SW480sub and HCT-15 cells invasion, SW480sub and HCT-15 cells were transiently (Fig. 5c). These results indicate the functional signifi- co-transfected with EGFR-specific siRNA and a cance of endosomal EGFR in promoting AKT and Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 11 of 20

Fig. 4 TMEM52B suppression promotes invasion and cell survival in an EGFR-dependent manner. (A) Immunoblot analysis of EGFR phosphorylation in cells transfected with shRNA specific to TMEM52B. (B) Cells were transfected with shRNA specific to TMEM52B for 48 h. Transfected cells were allowed to invade Matrigel in the absence or presence of EGF (10 ng/ml) for 48 h. Cell invasion was determined by calculating the cell-stained area relative to the total area using ImageJ software. (C-E) Cells were co-transfected with siRNA specific to EGFR, and shRNA specific to TMEM52B for 48. (C) Transfected cells were lysed for immunoblot analysis. Densitometric quantification was performed on the immunoblots using GAPDH as a loading control. Transfected cells were subjected to invasion (D) and cell survival (E) assays as described in Fig. 2A and B. All determinations were performed in three independent experiments. Values represent mean ± SD. *P < 0.05 compared with shControl + siControl; §P < 0.05 compared with shTMEM52B + siControl

ERK1/2 downstream signaling. Consistent with the re- Together, our results indicate that TMEM52B sup- sult of Fig. 5c, inhibition of EGFR endocytosis signifi- pression promotes the internalization and sustained ac- cantly reduced TMEM52B suppression-mediated cell cumulation of EGFR in the endosomal compartment in survival (Fig. 5d) and invasion (Fig. 5e). response to EGF treatment, resulting in enhanced Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 12 of 20

Fig. 5 TMEM52B suppression promotes phosphorylation and internalization of EGFR and downstream signaling. (A) Cells were transfected with shRNA specific to TMEM52B for 48 h and treated with EGF (10 ng/ml) for the indicated time prior to lysis for immunoblot analysis. Densitometric quantification was performed on the immunoblots; phosphorylated proteins were normalized against the corresponding total protein. (B) SW480sub cells were transfected with shRNA for 48 h and treated with EGF (100 ng/ml) for the indicated times. EGFR, EEA1, and DAPI were visualized; green for EEA1, red for EGFR, and blue for DAPI staining. Scale bar, 20 μm. Co-localization of EGFR and EEA1 was quantitated by calculating Pearson’s correlation coefficient using ImageJ software. Values represent mean ± SD. *P < 0.05. (C) Cells transfected with shRNA specific to TMEM52B for 42 h were treated with dynasore (160 μM) or DMSO as a vehicle control for 6 h and then treated with EGF (10 ng/ml) for 5 min prior to lysis for immunoblotting. (D, E) Cells were transfected with shRNA specific to TMEM52B for 48 h. Cell survival (D) and invasion (E) assays were performed as described in Fig. 2A and B in the presence of dynasore (5 μMor10μM) and EGF (10 ng/ml). Cell invasion was determined by calculating the cell-stained area relative to the total area using ImageJ software. All determinations were performed in three independent experiments. Values represent mean ± SD. *P < 0.05 compared with shControl + DMSO; §P < 0.05 compared with shTMEM52B + DMSO activation of EGFR and its downstream signaling media- viability mediated by TMEM52B overexpression. In tors AKT and ERK1/2 to drive cancer cell survival and addition, enforced expression of TMEM52B isoforms in invasion. HCT-116 cells, which display little to no TMEM52B expression, reduced phosphorylation of EGFR, AKT, JNK, TMEM52B overexpression suppressed invasion and ERK1/2, and c-Jun along with reduced cancer cell invasion survival of cancer cells and reduced phosphorylation of and survival (Supplementary Fig. S5), confirming the ef- EGFR, MAPKs, and AKT fects of the TMEM52B isoforms. On the other hand, We next explored whether TMEM52B overexpression TMEM52B overexpression did not substantially elevate E- suppressed cancer cell invasion and survival. SW480 cells cadherin expression in SW480 and HCT-116 cells (Fig. 6c were transiently transfected with TMEM52B-expressing and Supplementary Fig. S5C). vectors for 48 h prior to invasion and cell survival assays. To examine the effect of TMEM52B overexpression Overexpression of both isoforms significantly reduced on EGFR activity and EGFR-mediated cellular functions, cancer cell invasion (Fig. 6a) and survival (Fig. 6b). Immu- HEK293E cells were co-transfected with TMEM52B- noblot analysis showed that overexpression of both iso- and EGFR-expressing vectors. The overexpression of forms reduced phosphorylation of EGFR, AKT, JNK, both isoforms significantly suppressed both basal and ERK1/2, and c-Jun (Fig. 6c). AP-1 transcriptional activity EGF-induced cell invasion (Fig. 6e) (Notably, EGFR in SW480 cells was also decreased in response to expres- overexpression alone resulted in enhanced EGFR phos- sion of both isoforms (Fig. 6d). However, stable phorylation and increased invasiveness without the TMEM52B-overexpressing SW480 cells were not success- addition of exogenous EGF (data not shown). Therefore, fully established in our study, possibly due to reduced cell the moderate effect observed after exogenous EGF Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 13 of 20

Fig. 6 TMEM52B overexpression suppresses cancer cell invasion and survival and reduces phosphorylation of EGFR, MAPKs, and AKT. (A-C) SW480 cells were transfected with a TMEM52B-expression vector for 48 h. Invasion (A) and cell survival (B) assays were performed as described in Fig. 2A and B, respectively. (C) Transfected cells were lysed for immunoblot analysis. Anti-myc was used to detect myc-tagged TMEM52B. (D) SW480 cells were co-transfected with a TMEM52B-expression vector and an AP-1 reporter plasmid for 48 h. AP-1 activity was determined by a reporter assay as described in Materials and Methods. (E, F) HEK293E cells were co-transfected with an EGFR-expression vector and a TMEM52B-expression vector for 48 h. (E) Transfected cells (2 × 104 cells) were allowed to invade Matrigel for 48 h in the absence or presence of EGF (10 ng/ml). Cell invasion was determined by calculating the cell-stained area relative to the total area using ImageJ software. All determinations were performed in three independent experiments. Values represent mean ± SD. *P < 0.05. (F) Transfected cells were lysed for immunoblot analysis. An anti-myc antibody was used to detect myc-tagged TMEM52B treatment may have been due to a certain level of EGFR phosphorylation of EGFR, AKT, ERK1/2, JNK, c-Jun overexpression-mediated EGFR activation). Immunoblot (moderately), and ATF-2, and these changes were atten- analysis showed that EGFR overexpression enhanced uated by both isoforms (Fig. 6f). These results suggest Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 14 of 20

that TMEM52B reduces EGFR activation and that of play a role in activation of EGFR. Immunoblot analysis downstream signaling pathways, leading to suppression showed that EGFR phosphorylation was enhanced in of cancer cell invasion and survival. HCT-15 cells following treatment with purified recombinant E-cadherin fragment protein compared to mock TMEM52B suppression reduced E-cadherin stability to protein or no treatment (Fig. 7i). The phosphorylation of enhance β-catenin activity and promoted the generation ERK1/2 and AKT was also enhanced by treatment with of soluble E-cadherin fragment that contributed to EGFR soluble E-cadherin fragment (Fig. 7i). activation Together, these results suggest that TMEM52B sup- We next explored the mechanism of reduced E-cadherin pression reduces E-cadherin expression, potentially expression following TMEM52B suppression. Immuno- through modulation of E-cadherin stability, and pro- fluorescence analysis showed that TMEM52B- motes the extracellular shedding of E-cadherin, leading suppressed HCT-15 cells displayed significantly less E- to increased EGFR phosphorylation. cadherin, whereas scrambled siRNA-transfected cells maintained E-cadherin expression at organized cell-to- cell junctions (Fig. 7a). We examined whether the Clinical significance of TMEM52B in human cancers TMEM52B suppression modulated E-cadherin protein We examined whether TMEM52B expression correlated stability. SW480 and HCT-15 cells were transiently with E-cadherin expression in various cancer cell lines transfected with TMEM52B-specific shRNA for 42 h and and in human cancers. Correlations were determined by then treated with MG132, a proteasome inhibitor, for 6 calculating Pearson’s correlation coefficients (r). The h before lysis. TMEM52B suppression reduced E- analysis of various cancer cell lines from Cancer Cell cadherin levels, and this effect was substantially reversed Line Encyclopedia (CCLE) data demonstrated that by MG132 treatment (Fig. 7b). Co-immunoprecipitation TMEM52B mRNA expression was positively correlated studies, using whole-cell lysates from HEK293 cells co- with E-cadherin mRNA expression (n = 1156, r = 0.21, overexpressing myc-tagged TMEM52B and E-cadherin, P = 5.03E− 13), whereas TMEM52B mRNA level was revealed that both isoforms co-precipitated with E- negatively correlated with vimentin mRNA (n = 1156, cadherin (Fig. 7c) (Of note, two bands were detected by r=− 0.16, P = 3.33E− 08) (Fig. 8a). Analysis of kidney anti-E-cadherin, which may be due to E-cadherin modi- renal clear cell carcinoma data (TCGA, Firehose Legacy) fication such as phosphorylation in these cells). These showed that TMEM52B mRNA expression was posi- results suggest that TMEM52B binds to E-cadherin and tively correlated with E-cadherin mRNA expression (n = thus enhances its stability. 534, r = 0.26, P = 5.09E− 10) and E-cadherin protein ex- In addition, suppression of TMEM52B reduced β- pression (n = 478 for protein data, r = 0.12, P = 0.00784) catenin phosphorylation at Ser33, 37/Thr41, but en- (Fig. 8b). hanced phosphorylation at Ser675 (in SW480 and HCT- We also found that high TMEM52B expression in lung 15 cells) and Ser552 (in SW480 cells) (Fig. 7d), resulting and breast cancer patients was significantly correlated in enhanced nuclear transport of β-catenin (Fig. 7e). with increased overall survival and relapse-free survival, Those changes of β-catenin phosphorylation at Ser33, respectively, when survival within previously published 37/Thr41, and Ser552 may be mediated directly or indir- data sets was analyzed using the KM-plotter. TMEM52B ectly (through GSK-3) by AKT [28, 29]. Consistently, a expression in liver cancer patients showed a tendency reporter assay revealed that transcriptional activity of β- for a positive correlation with increased relapse-free sur- catenin was substantially increased by TMEM52B sup- vival (Supplementary Fig. S6A-C). We also found that pression in HCT-15 cells (Fig. 7f). On the other hand, kidney cancer patients with tumors that highly expressed TMEM52B suppression moderately reduced E-cadherin TMEM52B (Z > 1.5) had a significantly better overall promoter activity in SW480 cells, but not in HCT-15 survival rate than the remaining patients (Z ≤ 1.5) (Sup- cells (Fig. 7g), suggesting that TMEM52B suppression plementary Fig. S6D) from the analysis of TCGA- may reduce E-cadherin expression at the transcriptional generated kidney cancer data (TCGA, Firehose Legacy). level in a cell- or context-dependent manner. Furthermore, high expression of TMEM52B in lung, Recently, it has been reported that soluble E-cadherin, breast, ovarian, liver, rectal, and thyroid cancer patients an extracellular proteolytic fragment of E-cadherin, is el- was significantly correlated with increased survival when evated in the urine or serum of cancer patients, and can only patients with tumors that expressed E-cadherin, at activate EGFR family members [6]. Immunoblot analysis relatively low levels (below the median), were included of conditioned medium revealed that an extracellular in the KM-plotter analysis (Fig. 8c), suggesting that domain fragment of E-cadherin (~ 80 kDa) was shed TMEM52B expression may compensate for relatively from TMEM52B-suppressed HCT-15 cells (Fig. 7h), so low levels of E-cadherin to suppress tumor progression, we examined whether this fragment of E-cadherin may resulting in better survival. Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 15 of 20

Fig. 7 (See legend on next page.) Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 16 of 20

(See figure on previous page.) Fig. 7 TMEM52B suppression reduces E-cadherin stability and promotes extracellular shedding of E-cadherin, resulting in EGFR phosphorylation. (A) HCT-15 cells transfected with shRNA specific to TMEM52B for 48 h were immunofluorescent stained for E-cadherin (red) and nuclei (blue). Differential interference contrast (DIC) images are also shown. Scale bar, 10 μm. The level of E-cadherin-staining was calculated based on the ratio of the E-cadherin-positive area to the total observation area using ImageJ software. At least 7 random fields were analyzed. (B) Cells transfected with shRNA specific to TMEM52B for 42 h were treated with MG132 (2 μM) for 6 h before lysis for immunoblot analysis. (C) Co- immunoprecipitation analysis of the interaction between TMEM52B and E-cadherin in HEK293E cells co-transfected with myc-tagged TMEM52B and E-cadherin. Whole-cell lysates were immunoprecipitated with anti-myc and analyzed by immunoblotting with anti-myc or anti-E-cadherin. (D, E) Cells were transfected with shRNA specific to TMEMT52B for 48 h. (D) Transfected cells were lysed for immunoblotting. (E) A cytosolic fraction (denoted by C), nuclear fraction (N), and plasma membrane-enriched fraction (M) were prepared from transfected HCT-15 cells for immunoblot analysis of β-catenin. GAPDH, histone H3, and EGFR were used as internal controls for the cytosolic, nuclear, and plasma membrane fractions, respectively. Densitometric quantification was performed on the immunoblots using GAPDH or a fractional internal control as a loading control. (F) Transcriptional activity of β-catenin was analyzed using the TOP/FOP reporter system. HCT-15 cells were co-transfected with shRNA specific to TMEM52B and reporter plasmid for 48 h. Firefly luciferase activity was normalized against the Renilla luciferase activity and fold increases in TOPFlash activity compared to FOPFlash activity were calculated. (G) Cells were co-transfected with shRNA specific to TMEM52B and an E- cadherin promoter (− 308/+ 41) reporter plasmid for 48 h. E-cadherin promoter reporter activity was then measured. All determinations were performed in three independent experiments. Values represent mean ± SD. *P < 0.05. (H) HCT-15 cells were transfected with TMEM52B-specific shRNA for 48 h and the conditioned media from transfected cells collected for an additional 48 h. Lysates (intact E-cadherin) from transfected cells, and conditioned media (soluble E-cadherin; denoted as sE-cad) from cells were analyzed by immunoblotting. (I) Cells were serum-starved for 24 h and then treated with recombinant extracellular E-cadherin protein fused with Fc for 8 h. Cells were also treated with 25 μg/ml recombinant Fc protein as a negative control. Lysates were prepared and analyzed by immunoblotting. (J) A schematic representation illustrating the pathways for TMEM52B action in human cancer cells. TMEM52B suppression promoted activation and internalization of EGFR with enhanced MAPK and AKT signaling activity, leading to enhanced cell survival and invasion. In addition, TMEM52B suppression reduced E-cadherin stability, likely due to a reduced association between it and E-cadherin, resulting in enhanced β-catenin transcriptional activity. TMEM52B suppression- mediated generation of a soluble E-cadherin fragment, at least partially, contributed to EGFR activation. EE, early endosome. Nucleus was not distinguished. Dashed arrows indicate translocation

Discussion [30]. Whether TMEM52B exerts positive or negative in- We report here that TMEM52B plays a role in cancer fluences by cancer type or in a context-dependent man- cell invasion and survival, in an EGFR-dependent manner needs to be further explored. Similarly, TMEM52B ner, to reduce tumor growth and early metastasis. function and its precise molecular basis requires further TMEM52B suppression activated EGFR and downstream exploration during both tumor development and in nor- MAPKs and AKT signaling. In addition, TMEM52B sup- mal tissues. pression promoted EGFR internalization and sustained TMEM52B appeared to be involved in the modula- EGFR activation and signaling. TMEM52B enhanced E- tion of EGFR trafficking, based on our observation that cadherin protein stability through binding and thus TMEM52B suppression promoted EGFR endocytosis blocking its proteasomal degradation and extracellular and enhanced its signaling. In general, the primary role shedding. TMEM52B suppression-mediated generation of endocytosis is thought to terminate activation and of a soluble E-cadherin fragment, at least partially, con- signaling of receptor tyrosine kinases, including EGFR. tributed to EGFR activation. These results provide the However, internalization of activated EGFR has also first evidence that TMEM52B has tumor suppressor-like been reported to enable specific signaling pathways activity and is a novel modulator of EGFR and E- from intracellular compartments, maintaining its sig- cadherin (Fig. 7j). The clinical data analysis showed that naling outcome [8]. As an example of the importance TMEM52B expression was positively correlated with the of trafficking regulation in tumorigenesis, Vav2, a Rho increased survival of patients with several cancer types. GTPase guanine nucleotide exchange factor, is known TMEM52B might therefore be exploited as part of a to regulate cell adhesion, spreading, motility, and prolif- novel strategy for cancer treatment. For example, full- eration in response to growth factor signaling. Vav2 is length or partial fragments of TMEM52B may be candi- reported to delay EGFR internalization and degrad- dates for an anti-cancer drug through gene therapy or ation, thereby enhancing EGFR, ERK, and AKT phos- administration of recombinant suppressor proteins or phorylation [31]. Another example is the prolyl small molecule mimetics. hydroxylase PHD3, a scaffolding protein associated with Consistent with our results, it has been recently re- the endocytic adaptor Eps15. This protein promotes the ported that the downregulation of TMEM52B correlated internalization of EGFR and attenuates signaling to re- with poor survival of renal cell carcinoma patients [10]. duce cell proliferation and survival [32]. In contrast, re- In contrast, TMEM52B is also associated with poor sur- duction of SPRY2 (sprouty homolog 2) expression vival of gastric cancer patients, and reported to promote causes hyperactivation of PI3K/AKT signaling to drive gastric cancer cell invasiveness and metastatic capacity prostate cancer cell proliferation and invasion by Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 17 of 20

Fig. 8 Clinical significance of TMEM52B in human cancers. (A) Scatter plots examining TMEM52B mRNA expression (x-axis) and E-cadherin (left) or vimentin (right) mRNA expression (y-axis) from CCLE data (Broad, 2019). (B) Scatter plots examining TMEM52B mRNA expression (x-axis) and E- cadherin mRNA (left) or E-cadherin protein (right) expression (y-axis) from kidney renal clear cell carcinoma data (TCGA, Firehose Legacy). Correlations were statistically analyzed using the Pearson test. Equations were automatically generated using the cBioPortal webpage tool. (C) TMEM52B expression was correlated with the survival of cancer patients with tumors expressing low E-cadherin levels (below the median) when analyzed using the KM-Plotter. Kaplan-Meier analysis showed the probability of relapse-free survival from breast (n = 882), overall survival from lung (n = 571), post-progression survival from ovarian (n = 191), relapse-free survival from liver (n = 158), overall survival from rectal (n = 82), and overall survival from thyroid (n = 251) cancer patients data in relation to TMEM52B mRNA expression. TMEM52B expression was stratified as high vs. low according to an auto-select best cutoff value, and survival plots based on previously published data sets were generated using http:// kmplot.com. P values were calculated by the Logrank test enhanced internalization of EGFR and sustained signal- positive or negative effects on signaling and tumorigen- ing at early endosomes in a PTEN-dependent manner esis, suggesting conflicting or complex roles of EGFR [33]. Therefore, EGFR endocytosis can result in either trafficking. Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 18 of 20

On the other hand, we observed that TMEM52B sup- addition to the involvement of soluble E-cadherin, this pression failed to substantially affect the interaction be- may be attributed to the high expression of EGFR in the tween c-Cbl and EGFR or ubiquitination of EGFR cancer cells used in this study, potentially because over- (Supplementary Fig. S7A). In addition, TMEM52B sup- expression of EGFR and its oncogenic mutations can pression did not substantially affect EGFR dimerization lead to spontaneous dimerization of EGFR, resulting in (Supplementary Fig. S7B). Together, these results sug- receptor activation [8]. Regulation of EGFR activation gest that TMEM52B suppression promotes cancer cell has also been attributed to its association with flotillin at invasion and survival mainly through EGFR activation the plasma membrane, in addition to its interaction with and internalization although we cannot completely rule Ca2+/calmodulin complexes at the cytosolic juxtamem- out the possibility that other receptors are involved in brane region of EGFR [8]. It is also possible that EGFR TMEM52B function. Further studies should explore the was phosphorylated by autocrine production of EGFR precise molecular mechanisms of TMEM52B ligands. suppression-enhanced EGFR endocytosis and its sus- EGFR is a target for an expanding class of anticancer tained signaling. therapies. EGFR-inhibiting agents such as EGFR tyrosine E-cadherin exists in two forms: a membrane-tethered kinase inhibitors and EGFR-blocking antibodies are ap- form and a soluble form. Proteolysis of intact E-cadherin proved cancer treatments for a variety of cancers, in- leads to the release of an extracellular domain fragment cluding non-small-cell lung cancer, pancreatic, breast, of E-cadherin into the tumor microenvironment. This and colorectal cancers. Due to innate or acquired resist- soluble E-cadherin is constitutively shed at low levels in ance to these EGFR inhibitors, there remains an unmet normal epithelial cells, but significantly elevated in pri- need for further identification of potential molecular tar- mary tumor sites and metastatic foci. Correlations be- gets/pathways to enhance the development of novel ef- tween decreased survival and elevated soluble E- fective therapeutics. It is possible that TMEM52B-based cadherin levels in serum, urine, and tumors have been agents (e.g., TMEM52B functional-mimetics) may en- reported [34, 35]. Soluble E-cadherin is reported to fa- hance the therapeutic benefit of EGFR-inhibiting agents, cilitate tumor cell invasion, proliferation, and survival [6, although this requires further investigation. 36], thus exerting pro-oncogenic effects. We have demonstrated that TMEM52B suppression can downregulate Conclusions E-cadherin both transcriptionally and via altered protein In summary, we demonstrate that TMEM52B is a novel stability. Downregulation of E-cadherin results in re- modulator of E-cadherin and EGFR activity. TMEM52B duced cell-cell adhesion, thereby promoting cell motility β suppression reduces E-cadherin stability to produce sol- and activation of -catenin, thus contributing to malig- uble E-cadherin and enhances the activation and intern- nancy. In addition, the generation of soluble E-cadherin α alization of EGFR and its downstream signaling activity, induced by TMEM52B suppression (likely through - leading to cancer cell invasion and survival. These re- secretase, MMPs or ADAMs proteolysis) [6, 35], resulted sults suggest that TMEM52B has tumor suppressor-like in EGFR activation. Soluble E-cadherin is reported to ac- activity. Clinically, high TMEM52B expression correlated tivate EGFR family members and insulin-like growth fac- with increased survival among patients with varying tor receptor 1 (IGF-1R) on cancer cells and participate types of cancer, implicating the use of TMEM52B as a in the progression of various cancer types (autocrine sig- potential prognostic biomarker. This study highlights naling) [6]. It is possible that the soluble E-cadherin gen- the novel functions and underlying mechanisms of erated by TMEM52B suppression may exert its pro- TMEM52B, suggesting its prominent role in tumor tumorigenic activity directly to cancer cells and stroma progression. cells in the tumor microenvironment, although it needs further investigation. For example, soluble E-cadherin is reported to activate KLRG1, an inhibitory receptor on Supplementary Information The online version contains supplementary material available at https://doi. immune effector cells, to attenuate the cytotoxic effect org/10.1186/s13046-021-01828-7. to cancer cells [6, 37]. Therefore, the downregulation of intact E-cadherin and the production of soluble E- Additional file 1. cadherin, mediated by TMEM52B suppression, may contribute to the accelerated malignancy aggressiveness and Abbreviations tumor progression through both autocrine and paracrine CCLE: Cancer Cell Line Encyclopedia; CTC: circulating tumor cell; mechanisms. EGFR: epidermal growth factor receptor; EMT: epithelial-mesenchymal We observed that TMEM52B suppression enhanced transition; ERK1/2: extracellular signal-regulated kinase 1/2; GFP: green fluorescent protein; HEK293E: human embryonic kidney 293E; JNK: c-Jun N- EGFR phosphorylation in the absence of exogenous EGF terminal kinase; qPCR: quantitative polymerase chain reaction; shRNA: short as well as following treatment with exogenous EGF. In hairpin RNA; siRNA: small interfering RNA; TCGA: The Cancer Genome Atlas Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 19 of 20

Acknowledgments 12. Kim S, Kang HY, Nam EH, Choi MS, Zhao XF, Hong CS, et al. TMPRSS4 We are grateful to Prof. J. W. Lee (Seoul National University) for helpful induces invasion and epithelial-mesenchymal transition through discussion. upregulation of integrin alpha5 and its signaling pathways. Carcinogenesis. 2010;31(4):597–606. Authors’ contributions 13. Ahn HM, Ryu J, Song JM, Lee Y, Kim HJ, Ko D, et al. Anti-cancer activity of YL (first author) and DK performed most experiments and interpreted data; novel TM4SF5-targeting antibodies through TM4SF5 neutralization and JY helped animal experiments and interpreted data; YL (fourth author) immune cell-mediated cytotoxicity. Theranostics. 2017;7(3):594–613. helped with resources and discussed data; YL (first author) and SK analyzed 14. Kim S, Ko D, Lee Y, Jang S, Lee Y, Lee IY, et al. Anti-cancer activity of the clinical datasets; SK designed the study, interpreted data, and wrote the novel 2-hydroxydiarylamide derivatives IMD-0354 and KRT1853 through manuscript; and all authors read and approved the final manuscript. suppression of cancer cell invasion, proliferation, and survival mediated by TMPRSS4. Sci Rep. 2019;9(1):10003. Funding 15. Nam EH, Lee Y, Park YK, Lee JW, Kim S. ZEB2 upregulates integrin alpha5 This study was supported by grants from the National Research Foundation expression through cooperation with Sp1 to induce invasion during of Korea (NRF-2020R1A2C1008796, 2017R1A2B4002244, and epithelial-mesenchymal transition of human cancer cells. Carcinogenesis. – 2014R1A2A1A11049414 to S. Kim), and KRIBB Research Initiative Program, 2012;33(3):563 71. Republic of Korea. 16. La Monica S, Galetti M, Alfieri RR, Cavazzoni A, Ardizzoni A, Tiseo M, et al. Everolimus restores gefitinib sensitivity in resistant non-small cell lung cancer cell lines. Biochem Pharmacol. 2009;78(5):460–8. Availability of data and materials 17. Bourcy M, Suarez-Carmona M, Lambert J, Francart ME, Schroeder H, The datasets used and/or analyzed during the current study are available Delierneux C, et al. Tissue factor induced by epithelial-Mesenchymal from the corresponding author on reasonable request. transition triggers a Procoagulant state that drives metastasis of circulating tumor cells. Cancer Res. 2016;76(14):4270–82. Ethics approval and consent to participate 18. Ko D, Kim S. Cooperation between ZEB2 and Sp1 promotes cancer cell All animal procedures were performed in accordance with the guidelines of survival and angiogenesis during metastasis through induction of survivin the Animal Care Committee at the Korea Research Institute of Bioscience and VEGF. Oncotarget. 2018;9(1):726–42. and Biotechnology (KRIBB) and with approval of the bioethics committee of 19. Alcoser SY, Kimmel DJ, Borgel SD, Carter JP, Dougherty KM, Hollingshead the KRIBB (KRIBB-AEC-19098). MG. Real-time PCR-based assay to quantify the relative amount of human and mouse tissue present in tumor xenografts. BMC Biotechnol. 2011;11: Consent for publication 124. Not applicable. 20. Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, et al. Integrative analysis of complex cancer genomics and clinical profiles using Competing interests the cBioPortal. Sci Signal. 2013;6(269):pl1. The authors declare that they have no competing interests. 21. Cerami E, Gao J, Dogrusoz U, Gross BE, Sumer SO, Aksoy BA, et al. The cBio cancer genomics portal: an open platform for exploring multidimensional Author details cancer genomics data. Cancer Discov. 2012;2(5):401–4. 1Immunotherapy Research Center, Korea Research Institute of Bioscience and 22. Ghandi M, Huang FW, Jane-Valbuena J, Kryukov GV, Lo CC, McDonald ER Biotechnology, Daejon, South Korea. 2Department of Chemistry, Korea 3rd, et al. Next-generation characterization of the Cancer cell line Advanced Institute of Science and Technology, Daejon, South Korea. encyclopedia. Nature. 2019;569(7757):503–8. 3Department of Functional Genomics, Korea University of Science and 23. Gyorffy B, Surowiak P, Budczies J, Lanczky A. Online survival analysis Technology, Daejon, South Korea. software to assess the prognostic value of biomarkers using transcriptomic data in non-small-cell lung cancer. PLoS One. 2013;8(12):e82241. Received: 6 October 2020 Accepted: 3 January 2021 24. Gyorffy B, Lanczky A, Eklund AC, Denkert C, Budczies J, Li Q, et al. An online survival analysis tool to rapidly assess the effect of 22,277 genes on breast cancer prognosis using microarray data of 1,809 patients. Breast Cancer Res – References Treat. 2010;123(3):725 31. 1. Chaffer CL, Weinberg RA. A perspective on cancer cell metastasis. Science. 25. Menyhart O, Nagy A, Gyorffy B. Determining consistent prognostic 2011;331(6024):1559–64. biomarkers of overall survival and vascular invasion in hepatocellular 2. Yang J, Weinberg RA. Epithelial-mesenchymal transition: at the crossroads of carcinoma. R Soc Open Sci. 2018;5(12):181006. development and tumor metastasis. Dev Cell. 2008;14(6):818–29. 26. Gyorffy B, Lanczky A, Szallasi Z. Implementing an online tool for genome-wide 3. Thiery JP, Acloque H, Huang RY, Nieto MA. Epithelial-mesenchymal validation of survival-associated biomarkers in ovarian-cancer using microarray transitions in development and disease. Cell. 2009;139(5):871–90. data from 1287 patients. Endocr Relat Cancer. 2012;19(2):197–208. 4. Thiery JP, Sleeman JP. Complex networks orchestrate epithelial- 27. Nagy A, Lanczky A, Menyhart O, Gyorffy B. Validation of miRNA prognostic mesenchymal transitions. Nat Rev Mol Cell Biol. 2006;7(2):131–42. power in hepatocellular carcinoma using expression data of independent 5. Lamouille S, Xu J, Derynck R. Molecular mechanisms of epithelial- datasets. Sci Rep. 2018;8(1):9227. mesenchymal transition. Nat Rev Mol Cell Biol. 2014;15(3):178–96. 28. Fang D, Hawke D, Zheng Y, Xia Y, Meisenhelder J, Nika H, et al. 6. Hu QP, Kuang JY, Yang QK, Bian XW, Yu SC. Beyond a tumor suppressor: Phosphorylation of beta-catenin by AKT promotes beta-catenin soluble E-cadherin promotes the progression of cancer. Int J Cancer. 2016; transcriptional activity. J Biol Chem. 2007;282(15):11221–9. 138(12):2804–12. 29. Manning BD, Toker A. AKT/PKB signaling: navigating the network. Cell. 2017; 7. Sigismund S, Avanzato D, Lanzetti L. Emerging functions of the EGFR in 169(3):381–405. cancer. Mol Oncol. 2018;12(1):3–20. 30. Zhang JX, He WL, Feng ZH, Chen DL, Gao Y, He Y, et al. A positive feedback 8. Tomas A, Futter CE, Eden ER. EGF receptor trafficking: consequences for loop consisting of C12orf59/NF-kappaB/CDH11 promotes gastric cancer signaling and cancer. Trends Cell Biol. 2014;24(1):26–34. invasion and metastasis. J Exp Clin Cancer Res. 2019;38(1):164. 9. Strausberg RL, Feingold EA, Grouse LH, Derge JG, Klausner RD, Collins FS, 31. Thalappilly S, Soubeyran P, Iovanna JL, Dusetti NJ. VAV2 regulates epidermal et al. Generation and initial analysis of more than 15,000 full-length human growth factor receptor endocytosis and degradation. Oncogene. 2010; and mouse cDNA sequences. Proc Natl Acad Sci U S A. 2002;99(26):16899– 29(17):2528–39. 903. 32. Garvalov BK, Foss F, Henze AT, Bethani I, Graf-Hochst S, Singh D, et al. PHD3 10. Xie J, Zhu C, Wu J, Li C, Luo L, Xia L, et al. Down-regulation of C12orf59 is regulates EGFR internalization and signalling in tumours. Nat Commun. associated with a poor prognosis and VHL mutations in renal cell 2014;5:5577. carcinoma. Oncotarget. 2016;7(6):6824–34. 33. Gao M, Patel R, Ahmad I, Fleming J, Edwards J, McCracken S, et al. SPRY2 11. Hsu SC, Hung MC. Characterization of a novel tripartite nuclear localization loss enhances ErbB trafficking and PI3K/AKT signalling to drive human and sequence in the EGFR family. J Biol Chem. 2007;282(14):10432–40. mouse prostate carcinogenesis. EMBO Mol Med. 2012;4(8):776–90. Lee et al. Journal of Experimental & Clinical Cancer Research (2021) 40:58 Page 20 of 20

34. Okugawa Y, Toiyama Y, Inoue Y, Iwata T, Fujikawa H, Saigusa S, et al. Clinical significance of serum soluble E-cadherin in colorectal carcinoma. J Surg Res. 2012;175(2):e67–73. 35. Grabowska MM, Day ML. Soluble E-cadherin: more than a symptom of disease. Front Biosci (Landmark Ed). 2012;17:1948–64. 36. David JM, Rajasekaran AK. Dishonorable discharge: the oncogenic roles of cleaved E-cadherin fragments. Cancer Res. 2012;72(12):2917–23. 37. Melis L, Van Praet L, Pircher H, Venken K, Elewaut D. Senescence marker killer cell lectin-like receptor G1 (KLRG1) contributes to TNF-alpha production by interaction with its soluble E-cadherin ligand in chronically inflamed joints. Ann Rheum Dis. 2014;73(6):1223–31.

Publisher’sNote Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.